Modelling vocabulary acquisition in spoken word recognition Gareth - - PowerPoint PPT Presentation

Modelling vocabulary acquisition in spoken word recognition

Gareth Gaskell University of York

Models of human spoken word recognition

Box & Arrow - Cohort; Marslen-Wilson & Welsh, 1979) IAC – TRACE (McClelland & Elman, 1986), Shortlist (Norris, 1993) ASR-hybrid (Scharenborg et al., 2003) ART – (Vitevitch & Luce., 1999) Backpropagation – DCM (Gaskell & Marslen-Wilson,

1997)

In these models many aspects of word learning are either impossible or severely downplayed

DCM (Gaskell & Marslen-Wilson, 1997)

Recurrent Links Speech stream Distributed lexical representation

Semantic Output Phonological Output Hidden Units Phonetic Feature Input

Vocabulary acquisition and the lexicon

Current models of word recognition do not address changes in organisation of the system

developmental changes changes in adult processing (e.g., phonemic, lexical)

Example:

acquisition of novel words and their impact on lexical organisation

Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Pit

Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Speech input:

/k/

Pit

Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Speech input:

/k/

Pit

Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Speech input:

/k/

Pit

Stick Captain Can Collapse Cathedral Bed Ear Man Cathartic Coat Bread Speech input:

/ki/

Pit

Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Speech input:

/kidr/

Pit

Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Speech input:

/kidrl/

Pit

Novel word learning

What if we teach people a novel spoken word: “cathedruke”? This should enter the competition process and slow down recognition of existing words

Cathedruke Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Pit

Cathedruke Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Speech input:

/k/

Pit

Cathedruke Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Speech input:

/k/

Pit

Cathedruke Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Speech input:

/k/

Pit

Cathedruke Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Speech input:

/ki/

Pit

Cathedruke Stick Captain Can Collapse Bed Man Cathedral Cathartic Coat Ear Bread Speech input:

/kidr/

Pit

Cathedruke Stick Captain Can Collapse Cathedral Bed Ear Man Cathartic Coat Bread Speech input:

/kidrl/

Pit

Lexical access as filling in the features

“C…” “Ca…”

Time

“Cath…” “Cathed…” “Cathedral” Semantic Units Phonological Units

Novel word learning (Gaskell & Dumay,

2003; Dumay, Gaskell & Feng, 2004) Teach people novel spoken word: “cathedruke” Immediate effect on explicit memory

people can recognise “cathedruke” easily

Delayed effect on lexical processing

after 24 hours people are slower to recognise cathedral latest data suggest consolidation during sleep is the key

Delayed lexical competition

Recognition Test

490 500 510 520 530 540 550 560 Day 1 Day 2 Day 8 Response time (ms)

Control Novel Competitor

Dual-speed lexicalisation

Mental Lexicon New Vocabulary Phonological Trace/ Episodic Memory??

“One-shot” learning Localist representation? Hippocampal? Interleaved learning Distributed representation? Neocortical Facilitated by sleep?

Project proposal

Computationally & neurally explicit model of acquisition and storage of spoken words

Behavioural research

more on time course and informational circumstances underlying lexicalisation and other aspects of learning

Neuroimaging

Investigate neural bases of immediate and delayed aspects of word learning

Computational modelling

connectionist and statistical modelling of above

Behavioural research

More on:

Role of time in lexicalisation Role of sleep in lexicalisation Is there a lower limit on degree of exposure to novel items? Stability of lexical representations

Increase synergies with developmental research

Neuroimaging

fMRI research

neural correlates of one-shot learning and lexicalisation involvement of sleep cross-referencing with other types of memory consolidation during sleep

MEG research

track timecourse and localisation of lexical competition for novel items using magnetic MMN (Pulvermüller et al., 2003)

Computational Modelling

Starting point: DCM (Gaskell & Marslen- Wilson, 1997)

biologically plausible implement dual-speed learning systems integrate with more sophisticated front end integrate with other aspects of plasticity (e.g., age of acquisition, segmental adaptation)

Scope of model

Address three bodies of research; so far as possible relating behavioural and neuroimaging data

“steady-state” data on lexical access/lexical competition in speech perception plasticity, vocabulary learning, lexicalisation developmental language data